Karl Bode: it's important for outlets to explain that when we talk about AI "posing a threat to humanity," we're talking about the rushed implementation of half-cooked automation killing your grandma with an auto-rejection of her medicare claims, not sentient industrialized planet-scouring murderbots (for now)

please do not openly discuss our unethical energy and water sucking pump and dump www.washingtonpost.com/technology/2...

I still tend to think framing half-cooked "AI" as a "threat to humanity" helps make LLMs sound more impressive than they are, AND conflates critics with legitimate environmental and application concerns with crackpots worried that these error-prone parrots will suddenly and magically become Skynet

it's important for outlets to explain that when we talk about AI "posing a threat to humanity," we're talking about the rushed implementation of half-cooked automation killing your grandma with an auto-rejection of her medicare claims, not sentient industrialized planet-scouring murderbots (for now)

UnitedHealth uses AI model with 90% error rate to deny care, lawsuit allegesarstechnica.com For the largest health insurer in the US, AI's error rate is like a feature, not a bug.

"my grandma was killed by badly automated paperwork" sounds much less sexy than "AI has the ability to become sentient at any second and unilaterally decide to kill everyone in a campaign of sustained planetary terror

There's been a lot of chatter about the government using AI to identify terrorists. 🙃

it's like the buttle/tuttle screw up from Terry Gilliam's Brazil, but at unfathomable scale

We're going to start seeing an uptick in wedding massacres.

www.cnn.com/2024/04/03/m... 🫠 already happening some places!

Israel is using artificial intelligence to help pick bombing targets in Gaza, report says | CNNwww.cnn.com The Israeli military has been using artificial intelligence to help identify bombing targets in Gaza, according to an investigation by +972 Magazine and Local Call, citing six Israeli intelligence off...

Yeah, I'm not watching the movie about the first statement.

yeah A24 still might be able to cobble together a sad indie about it

Well, if A24 is producing it...

There's probably a combination of the two which will actually be a concern.

yeah eventually, but we're talking about the now and a fully sentient AI is still far away

tbf, health insurance companies don’t need help rejecting claims or disallowing treatment. This will just lower thecocst for them.

well it automates it badly at impossible new scale, AND it creates a potential and legal firewall for unethical behavior ("the system" did it, we were just trusting the software)

That's it. Trying to remove accountability.

they not being held accountable now anyway. they do whatever they want.

not to be a debbie downer doomer but it's going to be much much worse now that the supreme court has eviscerated regulatory independence they'll all be so paralyzed by the threat of corrupt judge rulings you won't even get these occasional good faith efforts we see now

It is rare and poor, but it does happen. Making it harder is the goal. www.atg.wa.gov/news/news-re...

AG Ferguson: Insurance company must repay every penny to Washingtonians it denied due to mental health discrimination | Washington Statewww.atg.wa.gov Lawsuit results in important reforms to the travel insurance industry to provide protections for individuals experiencing severe mental health events OLYMPIA — Attorney General Bob Ferguson announced ...

If there's an upside, it's that this removes the excuse that doing away with health insurance altogether (in favor of universal Healthcare) would destroy jobs

📌

*taps the sign*

If we deal with the bad paper-pushing "AI" then maybe we have murderbot "AI." As a treat.

Kill AI

+Present, ongoing, harms are inconvenient admissions of guilt for investors eager to live in the land of speculative capital. If improving health was the real goal, the possibility would be mute because rationed private healthcare decision making wouldn't even be a thing.

I thought you were going to comment on their energy consumption, which is doing tremendous damage. Just so a chatbot can give me a recipe from my 3 ingredients, or summarize an article I'm too lazy to read. Something I used to be able to Google.

When I worked in healthcare processing insurance claims I ALWAYS called United Healthcare directly and spoke to a human because I didn’t trust their online auth processor. Even if it says a code passed online they will deny a claim after the fact and then tell the patient we never called.

There are people in healthcare that would totally agree. I know where this company's dead bodies are buried

Yes, you do

Intended to thin the herd making it easier for the murderbots.

clever bastards

What would Bender do?

📌

My blog bit on the glamorizing of the robot apocalypse massmiss.blog/2024/02/03/t...

TESCREALs: AGI, LLMs...… the whole toxic alphabet soupmassmiss.blog The reality of Artificial General Intelligence, Large Language Models, etc., compared to TESCREALism, the philosophy of billionaire tech-bros.

📌

I know someone who was an executive there. He was & still is a deplorable man, a Disgusting example of a male human being

Post